From 8753ad8b27140489a36699ed58ae78e02bf4b24c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andreas=20Kusalananda=20K=C3=A4h=C3=A4ri?=
<ak4@sanger.ac.uk>
Date: Wed, 9 Feb 2011 14:48:04 +0000
Subject: [PATCH] Add code to parse and store synonyms from the OBO file. Also
only get term definitions from within "...".
---
.../ontology/scripts/load_OBO_file.pl | 19 ++++++++++++++++---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/misc-scripts/ontology/scripts/load_OBO_file.pl b/misc-scripts/ontology/scripts/load_OBO_file.pl
index 6dfb25e45a..048f1a00ae 100755
--- a/misc-scripts/ontology/scripts/load_OBO_file.pl
+++ b/misc-scripts/ontology/scripts/load_OBO_file.pl
@@ -135,14 +135,17 @@ sub write_term {
print("Writing to 'term' table...\n");
- $dbh->do("LOCK TABLE term WRITE");
+ $dbh->do("LOCK TABLE term, synonym WRITE");
my $statement =
"INSERT INTO term "
. "(ontology_id, subsets, accession, name, definition) "
. "VALUES (?,?,?,?,?)";
- my $sth = $dbh->prepare($statement);
+ my $syn_stmt = "INSERT INTO synonym (term_id, name) VALUES (?,?)";
+
+ my $sth = $dbh->prepare($statement);
+ my $syn_sth = $dbh->prepare($syn_stmt);
my $id;
my $count = 0;
@@ -181,6 +184,13 @@ sub write_term {
}
$term->{'id'} = $id;
+ foreach my $syn ( @{ $term->{'synonyms'} } ) {
+ $syn_sth->bind_param( 1, $id, SQL_INTEGER );
+ $syn_sth->bind_param( 2, $syn, SQL_VARCHAR );
+
+ $syn_sth->execute();
+ }
+
++$count;
} ## end foreach my $accession ( sort...)
alarm(0);
@@ -432,7 +442,7 @@ EOT
} elsif ( $type eq 'namespace' ) {
$term{'namespace'} = $data;
} elsif ( $type eq 'def' ) {
- $term{'definition'} = $data;
+ ( $term{'definition'} ) = ( $data =~ /^"(.*)"/ );
} elsif ( $type eq 'is_a' ) {
my ($parent_acc) = $data =~ /(\S+)/;
push( @{ $term{'parents'}{'is_a'} }, $parent_acc );
@@ -443,6 +453,9 @@ EOT
if ( $data eq 'true' ) { $state = 'clear' }
} elsif ( $type eq 'subset' ) {
push( @{ $term{'subsets'} }, $data );
+ } elsif ( $type eq 'synonym' ) {
+ my ($synonym) = ( $data =~ /^"(.*)"/ );
+ push( @{ $term{'synonyms'} }, $synonym );
}
} ## end elsif ( $line =~ /^(\w+): (.+)$/)
--
GitLab